Introduction to Risk Management

Introduction to Data Science in Python

FinTech Foundations and Overview

Introduction to Self-Driving Cars

A quick tour on Big Data and Business Intelligence

Fundamentals of Finance

Foundations of Data Science

AFAD Gönüllüsü

Kan Bağışı Destek Gönüllüsü

Lisans Derecesi

Engineered an end-to-end AI platform processing ~1 million rows of NYC Taxi & Rideshare data to optimize urban transportation networks.

Key Business & Technical Achievements:
🧠 High-Accuracy Prediction: Built Machine Learning models (Random Forest) achieving 90% accuracy (R²: 0.899) in hourly demand forecasting.
📉 Operational Optimization: Reduced trip duration estimation error to ~3 minutes (MAE) using XGBoost, improving ETA reliability.
💸 Revenue Strategy: Identified route and hotspot optimization strategies capable of increasing driver revenue by ~20%.
🚦 Dynamic Simulation: Modeled congestion pricing and dynamic pricing surges to simulate market behavior during peak hours.

Technical Stack & Implementation:
• Developed an automated ETL pipeline handling massive datasets (1GB+) using PyArrow and Pandas for memory-efficient processing.
• Created a modern, interactive dashboard using React.js and Tailwind CSS to visualize KPIs like market share, traffic density, and regional deman

Data Scientist (Urban Mobility & AI)

Developed "Bio-GPT," a Transformer-based Deep Learning model to predict cancer cell states using single-cell RNA sequencing (scRNA-seq) data, bridging the gap between AI and precision medicine.

Scientific & Technical Achievements:
🧬 High-Precision Diagnosis: Achieved 98.70% accuracy in classifying cell states (Healthy, Early/Advanced Cancer, Apoptotic) on unseen test data.
🔍 100% Interpretability: Utilized Self-Attention mechanisms to identify critical gene biomarkers (e.g., Gene_143, Gene_647) for potential drug discovery.
⚡ Process Acceleration: Automated scRNA-seq data preprocessing pipelines, reducing analysis time by ~70% compared to manual methods.
📊 Advanced Visualization: Implemented UMAP and PCA for dimensionality reduction, proving distinct clustering of cancer states in the gene embedding space.

Technical Stack:
• Architecture: PyTorch, Hugging Face Transformers, Self-Attention Mechanisms.
• Bioinformatics: Scanpy, AnnData for single-cell analysis.
• Data Science: Sciki

AI Research Engineer (Bioinformatics & Deep Learning)

Developed an advanced Predictive Customer Lifetime Value (CLV) Engine for e-commerce, utilizing probabilistic models and machine learning to optimize marketing strategies.

Key Business Impacts:
🚀 Revenue Protection: Identified $225K+ in saveable revenue by detecting "At Risk" customers.
💰 Marketing Efficiency: Achieved 50% reduction in marketing costs by focusing on the top 50.5% of customers who generate 64.7% of revenue.
📈 Forecast Accuracy: Improved prediction accuracy by 35% using BG/NBD and Pareto/NBD models compared to traditional methods.
🎯 ROI Boost: Projected 15-25% total revenue increase through targeted VIP segmentation.

Technical Implementation:
• Applied RFM Analysis (Recency, Frequency, Monetary) on 100,000+ transactions and 10,000 customers.
• Implemented K-Means Clustering to discover 3 natural customer segments driven by purchasing behavior.
• Utilized Python libraries (Lifetimes, Scikit-learn, Pandas) for probabilistic modeling and data processing.
• Created a 3

Data Scientist (Predictive Analytics & ML)

Data Scientist | Python, NLP & Machine Learning | Building AI for Healthcare & E-Commerce (BioGPT & CLV Analysis) | Comp

Know about Ege Koçaslan's work experience, education, bootcamps, certifications, and much more. Ege Koçaslan